Target center extraction method based on joint calibration of camera and radar

Compared with the traditional feature point extraction, an improved algorithm is proposed to accurately extract the center point coordinates of the target with deflection angle. On the one hand, the algorithm extracts the pixel center point coordinates through the improved fitting ellipse method. On the other hand, the joint calibration is carried out through the pose design of the camera and the radar to calculate the required parameters and improve the accuracy of the algorithm. The error comparison experiment is designed. The experimental results show that, first of all, for the average root mean square error, the error after joint calibration is reduced by about 17.01%∼ 20.54%, indicating that the deviation between the measured value and the real value is reduced, which is closer to the actual model. Secondly, for the average re-projection error, the error after joint calibration is reduced by about 17.22%∼17.89%, indicating that the results of three-dimensional reconstruction are more accurate than the original. Finally, the accuracy of the joint calibration results of the camera and radar is verified by the error histogram.